In recent years, machine learning models have emerged as potent tools for prediction studies, particularly when dealing with sequential data. This research delves into the impact of Convolutional Long Short-Term Memory (Conv-LSTM) model parameters on the accuracy of predictions using the Normalized Difference Salinity Index (NDSI) time series. The Conv-LSTM model architecture constitutes an evolution of the Long Short-Term Memory (LSTM) model, engineered to capture both spatial and temporal dependencies in sequences. Unlike traditional LSTMs, Conv-LSTMs embrace the spatial structure of input data by coupling LSTM units with convolutional layers. This synergistic design allows Conv-LSTMs to extract and integrate features across both time and space, making them adept at analyzing complex time series data with spatial correlations, such as satellite images or environmental measurements. The NDSI is an indicator of salinity in water bodies and coastal regions, making it a valuable tool for environmental monitoring. Utilizing Landsat-8 data of Tangier city of Morocco collected between 2015 and 2022, the NDSI time series dataset was constructed, forming the foundation for evaluating prediction accuracy. In this study, Conv-LSTM model was configured with three pivotal parameters: i) the number of “filters” in the main layer, ii) the number of “neurons” in the fully connected layer, and iii) the number of training “epochs”. The NDSI time series data were employed to train and evaluate the model, with prediction accuracy assessed using the coefficient of determination (R²) metric. The results uncover substantial insights into the relationship between Conv-LSTM model parameters and prediction accuracy for NDSI analysis. When considering a high number of epochs (i.e., epochs=100), the prediction accuracy remained relatively consistent at 97% across varying values of filters and neurons. This suggests that rendering the number of epochs beyond a certain point less influential on accuracy improvements. In the context of medium number of epochs (i.e., epochs=50), the observed accuracy variations were more pronounced. Notably, the accuracy was influenced by the number of filters in the main layer. Specifically, when filters numbered between 10 and 100, accuracy remained below 60%. However, with a rise in filter count, accuracy exhibited an upward trend, ultimately plateauing at 96%. In contrast, for a low epoch count (i.e., epochs=10), the initial accuracy was negative. However, this was addressed by introducing an extensive number of filters in the main layer, reaching up to 10,000. The infusion of this high filter count yielded positive accuracy outcomes reaching more than 60%, indicating that a substantial filter count compensated for the limited training epochs.

Rivers are the dynamic water resources available on the earth surface. River morphology focuses on water courses' shape and pathways, and how they change over time. In this study, we have focused on the Narayani river using remote sensing techniques to understand the dynamics on river channel shifts . Two different time intervals of 15 years were taken to understand how rivers are changing. We used Landsat 5, Landsat 7, and Landsat8 for rivers images of 1990, 2005, and 2020 respectively. Supervised classification of satellite images was done using ArcGIS Pro 3.0.3, and images are classified into water and non-water bodies. For quantification of river shifting, we adopted two approaches: linear displacement and area change method. In the linear displacement method, we bisected the river with an imaginary straight line along the river length as per latest river flowpath of 2020. Then, we selected the sample points at the fixed distance to measure the width of the river and avoid biases. Finally, we measured river width within those prespecified sample points across the river in each location for all water courses. Thus, river widths are measured in a straight line across the river, averaged them and quantified the change in water course in the span of each 15 years. Secondly, in the area change method, we used image differencing techniques from later to former to see changes in the water course.

Interestingly, the water course of Narayani river increased from 1990 to 2005 but it decreased sharply in 2020. The shifting of rivers was intense in forest areas compared to other land uses. The river width ranges from 80 meters to 302 meters in 2020. Regarding the area, the water course decreased by 2.3 acres and increased by 3.8 acres in 1990-2005. Similarly, the water course decreased by 4.14 acres and increased by 2.25 acres in 2020. Monsoon rain leading to annual flooding is the primary cause for this change in river course. Climate change can be related as a secondary cause for these floods and river shifting because of melting glaciers and increasing water discharge. Apart from this, upland land use change can cause change in water course of Narayani river. This study can be a baseline for researchers to correlate the shifting in river course and its factors; while planners and policy makers can come up with a strategic plan to halt down this process of river change and maintain this natural repository of freshwater intact in the long-run.

Tropical protected and reserved forests are the most active and diversified ecosystem that plays a significant role in maintaining the ecological balance. These forests are experiencing the problem of extensive forest deforestation due to illegal encroachment by the surrounding communities. Assessing forest cover dynamics and its health assessment plays a significant contribution to the management of the forest biodiversity and ecosystem. The present study aims to assess the forest dynamics using multi-temporal satellite data from 1990-2022 (i.e., 10 years intervals of satellite images from Landsat-5 and Landsat-8). The study has been conducted in Behali Reserve Forest (BRF) located in the northeastern Himalayas. Furthermore, we assessed its health conditions based on biophysical and biochemical parameters derived from Sentinel-2A satellite images and near-proximal sensors. The key findings indicate that over the span of 1990-2022, there is considerable forest cover losses of 15% during the 1990s, with a gain of 1.3% during the 2000s and 0.01 % during the 2010s. The net change in forest dynamics showed about 25.2 km² (17.3%) areas undergone deforestation while 3.7 km² (2.5%) areas expanded under afforestation in the last three decades (1990-2022).

Health assessment was performed by using LAI and leaf chlorophyll which serves as a key indicator of leaf canopy density and photosynthetic pigment, respectively. The key findings related to health assessment indicate that LAI ranged from 1 to 5.5 and the healthy dense forests showed LAI ≥ 4.5. The Normalized Area Over Reflectance Curve (NAOC) index based leaf chlorophyll content of dense forests showed that chlorophyll ranged between 30 and 45 μg/cm². The leaf chlorophyll content from satellite and field-based measurements exhibited a coefficient of determinants (R²) of 0.88, indicating strong relationships. We can conclude that this study helps to exemplify the potential of remote sensing techniques to evaluate the dynamics of tropical forests and their biophysical and biochemical properties. It provides critical information about forest dynamics and its health conditions, which are useful for forest conservation and management.

Monitoring condition and quality indicators within river water systems has emerged as a pressing priority due to the decline in water quality. This decline is primarily a result of improper disposal of household waste and the discharge of partially treated or untreated sewage and industrial effluents into the neighboring water bodies connected to the river systems. Conventional techniques for assessing water quality are costly and intricate and demand significant labor. Instead of relying on conventional field sampling methods, the utilization of cost-effective satellite imagery holds significant potential for the foreseeable future. This current study focuses on employing LANDSAT-8 OLI imagery to assess the Brahmani River's water quality parameters over 5 years, from 2017 to 2021. The selected water quality parameters encompass Biochemical Oxygen Demand (BOD), Dissolved Oxygen (DO), pH levels, Total Coliform (TC), and Fecal Coliform (FC). To determine the reflective values of all bands relevant to the aforementioned water quality parameters at their respective sampling sites, the Sentinel Application Platform (SNAP) tool was employed. Valuable insights were gained by establishing linear correlations between the water quality parameters and the reflective values of the bands and band ratios. Notably, the Pearson Correlation Coefficients revealed strong associations, with values of 0.892 for pH and B3/B7, 0.746 for DO and B1/B2, 0.814 for BOD and B7/B6, and 0.875 for FC and B6/B3. However, no robust correlation was observed between TC and any of the band ratios. The coefficients of determination for pH, DO, BOD, FC, and the corresponding band ratios (B3/B7, B1/B2, B7/B6, B6/B3) were calculated to be 0.796, 0.765, 0.772, and 0.766, respectively. Utilizing the outcomes of the linear regression analysis for pH, DO, BOD, and FC, predictions were made for the water quality parameters in the years 2020 and 2021. Impressively, a strong alignment was evident between the projected and observed values. This led to the conclusion that the established equations exhibited a noteworthy capacity to accurately predict the water quality parameters for the Brahmani River.

The spatial distribution of rice fields in the Mazandaran province is of utmost importance for understanding various crucial aspects such as food security, water usage, greenhouse gas emissions, and disease transmission. Agricultural irrigation plays a significant role in expanding crop lands as well. However, the availability of limited information regarding cropland areas, particularly in the Mazandaran province, is a challenge. To address this issue, we have employed a pixel-based paddy rice mapping (PPPM) algorithm to generate accurate ground truth data. This algorithm identifies flooding signals during the rice transplanting phase, providing the necessary ground truth data. We have also cross-checked the transplanting start and end times with Mazandaran's agricultural calendar for further validation. Furthermore, we have proposed a novel method that utilizes machine learning and time series Sentinel-1 and Sentinel-2 images in the Google Earth Engine system to differentiate rice fields from other types of crop lands. The proposed method involves several steps. Firstly, we have calculated various bands and essential indicators such as blue, green, red, red edge 1/2/3, NIR, nNIR bands, as well as NDVI, LSWI, DVI, RVI, WDRVI, SAVI, EVI, VARIGREEN, and GNDVI from Sentinel-2 images. Secondly, we have utilized multi-collinearity analysis to obtain the optimal indexes and bands, which include NIR, Red Edge 3, NDVI, LSWI, VARIgreen, DVI, and GNDVI. In the third step, we have generated monthly composites of the optimal indices and bands from March to August. Subsequently, we have employed the Random Forest classification algorithm to classify the study area into six classes: water, crop land, urban, forest, outcrop, and range land. Finally, we have used the Radar Vegetation Index extracted from Sentinel-1 to accurately separate the rice fields from other types of crop lands. Our approach arrived impressive results with a high accuracy and kappa coefficient of 99% and 98%, respectively. This information is crucial for policymakers and researchers as it enables them to make informed decisions concerning food security, water usage, greenhouse gas emissions, and disease transmission in the Mazandaran province. Additionally, it provides valuable insights into the expansion of crop lands through agricultural irrigation. By leveraging machine learning and satellite imagery, we are able to generate accurate and reliable information about cropland areas, facilitating the development of effective strategies for sustainable agriculture and food security.

GNSS interferometric reflectometry (GNSS-IR) can be considered as another remote sensing technique for continuous and local monitoring of soil moisture content, which can be performed in various weather conditions such as rainy and cloudy conditions, as well as different lighting conditions such as day and night. In GNSS-IR Changes in soil moisture result in changes in the signal-to-noise ratio (SNR) component of reflected signals. By analyzing reflected signals, useful information about the reflector can be obtained. SNR is highly dependent on soil moisture. As the soil moisture content increases, the dielectric constant of the soil increases, which causes the reflected signals to have higher amplitudes and higher SNR. Conversely, as the soil moisture content decreases, the reflected signals have lower amplitudes and lower SNR. Therefore, analyzing the SNR of the reflected signals can provide useful information about the soil moisture content.

In this study, data from station P038 in the New Mexico region is used, where multipath signals are used to estimate soil moisture changes over a four-year period from 2017 to 2020. This research has four main steps. In the first step, appropriate satellite tracks with elevation angle between 5 to 30 degrees are selected. SNR data are generated from RINEX files. Then, the initial reflection height is estimated for each path. The phase is obtained for each satellite on each day through several stages. Next, SNR metrics are calculated, and finally, vegetation cover effects are mitigated and removed and the result is converted to volumetric water content.

According to the estimates, the volumetric water content in 2017 was 8.88, which increased to 11.74 in 2018, then slightly decreased to 10.88 in 2019 and finally increased to 12.49 in 2020. In this article, the effectiveness of the LSTM neural network model in predicting the time series of volumetric soil moisture obtained from GNSS-IR signals is evaluated. This prediction will help farmers to prepare their irrigation schedules more efficiently.

The LSTM neural network can maintain its content over a long period of time and essentially remember previous information. Gates are also vectors with values between zero and one that determine how old information should be progressed and new information should be added. Generally, one means to pass and zero means to discard information. The input gate specifies which parts of the input data and to what extent they should be added to the memory content. The forget gate determines which parts of the memory content should be removed. The output gate also determines which part of the hidden state content should contain the memory content.

The model is trained using 80% of observations. By updating the network status with observed values instead of predicted values, the RMS error decreased from 0.09 to 0.05, and the predictions became more accurate. Investigations have shown that performing GNSS observations produces more homogeneous reflective effects around the antenna. Therefore, in order to increase the accuracy and quality of the results, it is suggested to use GNSS interferometric reflectometry instead of just GPS interferometric reflectometry.

Buildings' classification using satellite images is gaining importance for several applications, such as damage assessment, resource allocation, and population estimation. This work focuses on two specific classification tasks: building-type classification (residential or non-residential) and building-damage assessment (damaged or not-damaged). These two tasks have become of great interest recently, especially considering current global conflicts and natural disasters, such as the Ukrainian conflict, the 2023 Turkey–Syria earthquake, and the recent tragic flood in Libya's Derna.

Existing building-type classification approaches combine optical satellite images with LIDAR data or street view images. We propose here to rely solely on RGB satellite images and follow a 2-stage deep learning-based approach, where buildings' footprints are extracted using a semantic segmentation model, followed by the classification of the segmented images. To the best of our knowledge, this is the first Building-Type Classification (BTC) method solely based on RGB aerial imagery. Optical imagery is cost-efficient, widely available, and can be tasked and acquired quickly following a military conflict or natural disaster.

Supervised deep learning algorithms rely on high-quality labeled datasets. Four Building-Damage Assessment (BDA) datasets are considered, analyzed, and compared to choose the appropriate one for our research. For the scope of this work, we define a damaged building as a partly or wholly demolished building that might result from armed conflicts, earthquakes, tornadoes, and hurricanes. On the other hand, due to the need for an appropriate residential/non-residential building-type classification dataset, we introduce a new high-resolution satellite imagery dataset called the Beirut Buildings Type Classification (BBTC) dataset.

We conduct experiments to select the best hyper-parameters and model architecture and propose an extra transfer learning stage that outperforms the classical method. Finally, we validate the performance of the proposed schemes, where BTC achieved a 94.8% accuracy using RexNet as a backbone, and BDA scored 97.3% accuracy using Focal loss and Adam optimizer. With the extra transfer learning stage, BDA improved in the last percentages of the accuracy and F1-score to reach 98.96% and 99.4%, respectively.

Forests are considered vital to earth and humanity, as they clean air, protect us from disasters, and enhance our well-being. However, nowadays forests are frequently exposed to different types of dangers, and fires are the major ones, as they drastically alter the land surface properties, including temperature and vegetation. In this research, we aim to assess the impact of fire on surface temperature and vegetation using the Land Surface Temperature (LST) index and the Normalized Difference Vegetation Index (NDVI), respectively. As a study case, we chose two zones of the “Bou Jedyane” forest (Mediterranean region, North of Morocco), one of which was totally affected by a fire occurred in July 2022, with an area of approximately 115 Km², and the other was used as a control. Landsat-8 images from January 2021 to August 2023 were used to extract the considered indices. Then, we computed their percentage of difference between the two zones to assess how the fire changed the temperature and vegetation in the burned area compared to the intact area before, during, and after the fire. Results revealed that fire significantly affected LST and NDVI in the region, as they shifted dramatically from percentage differences ranging between 1% and 7% in 2021 to 27% and 59% for LST and NDVI, respectively, right after the fire outbreak. In the year 2023, we got a decreasing percentage of differences in both LST and NDVI, indicating that the forest is recovering over time. These results demonstrate the impact of fire on two significant elements, the land surface temperature and the vegetation, as well as the forest potential for natural regeneration.

The escalating environmental concern of salinity has far-reaching impacts on the global community. The detrimental effects of salinity significantly degrade soil fertility, resulting in economic losses in agriculture and posing a threat to food security. Monitoring and analysing changes in soil salinity over time are essential tasks for effectively devising strategies in natural resource management for the future. Numerous models based on multispectral and dielectric techniques have been formulated to estimate soil salinity utilizing data from Landsat and Sentinel satellites. Palacode Taluk, located in the Dharmapuri district of Tamil Nadu, stands out as an area severely affected by natural salinity-related factors. Consequently, there is an urgent necessity to pinpoint the most accurate assessment method for salinity in this region, thereby improving the well-being of farmers, their livelihoods, and the overall ecosystem. This study aims to investigate and compare the effectiveness of different salinity models derived from Sentinel-1, Sentinel-2, and Landsat 8 OLI satellites. Various soil salinity indices, acquired through different combinations of visible and infrared bands from Sentinel 2A, along with a modified salinity index developed from Landsat data, were scrutinized to explore the potential of multispectral models. Additionally, the study evaluated the performance of dielectric models by examining the DSDM and Hallikainen model. Statistical analysis and validation were conducted through linear regression analyses to establish correlations between on-site measurements and satellite-based models. The results suggest that dielectric models produce better results for the specified study area in comparison to multispectral models. Specifically, the Hallikainen model performed the best, demonstrating a stronger correlation with on-site measurements. Within the multispectral models, the modified salinity index generated from Landsat data exhibited a greater correlation with actual soil salinity measurements than the models utilizing various combinations of bands from sentinel 2A data.

The northern coastal of Oman along with the Strait of Hormuz is an incredibly important gateway since it is a key water access for the transit of oil and gas around the world. Due to its important geographical location, any oil spills in this area might have serious impacts on the world economy and energy markets in addition to the marine habitat. Considering how crucial it is to actively monitor and respond to oil leak accidents, this study was carried out expressly to investigate the likelihood of them occurring between April and May 2020. In this research study, we investigated such oil spills used SNAP (Sentinel Application Platform), a robust software tool that is frequently used in remote sensing applications. Further, satellite imagery data were made available through the Copernicus Open Access Hub, a website run by the European Space Agency (ESA). This study specifically made use of the Sentinel-1 satellite's radar imagery data, which provide high-quality and precise data for tracking oil spills.